JorgeB

Either way will work, use the most convenient for you.

I tried both and could not find any performance difference with Unraid, still ended up using the LSI firmware just in case.

Should be plug n play, didn’t need to disable bios on mine, it needs latest firmware for >2tb support.

Yep

I did one recently, not because of a failure but for upgrading a 3tb drive, duration was similar to a parity check/sync. Event: unRAID Data rebuild: Subject: Notice [TOWER7] - Data rebuild: finished (0 errors) Description: Duration: 14 hours, 55 minutes, 14 seconds. Average speed: 149.0 MB/sec Importance: normal

Add it after append, it should work, maybe it works even for other marvell chipsets, let us know.

This is how it always worked for me.

Either will work, note however that XFS only supports single cache, for a pool you need BTRFS.

Install Dynamix system stats plugin, start building/checking one and look at the CPU load, if low start another one and so on. With MD5 I can do four at a time in my fastest servers, 2.5 to 2.7Ghz Dual Core Sandy and Ivy Brige CPUs, my HPs N54L microservers on the other hand can only handle two at a time.

Go to tools and click new config, it will create disk.cfg

Fstrim won't work, you'll get fstrim: /mnt/cache: FITRIM ioctl failed: Input/output error

First confirm your SSD has TRIM support: hdparm -I /dev/sdX Look for: * Data Set Management TRIM supported (limit 8 blocks) If it does your possibly using an HBA that doesn't support FSTRIM, e.g., LSI9211-8i.

Although you’re not using parity, you can’t replace one data disk with another, you have to go to tools and do a new config, then just re-assign all disks you want to use now.

Sounds good.

You can ignore the command timeout on value on Seagates, reported uncorrected above 0 on the other hand usually means a higher probability of failure, note however that rebuilding disk6 to a new disk won’t fix the filesystem corruption, but that’s what I would do, so you’ll still have old disk6 to try and recover data if the reiserfsck can’t recover the rebuilded disk.

I believe this happens when the check for plugins detects a new version, but in the meantime a newer one is released, so it’s trying to download one release and using md5 from a different one, checking for updates in the plugins page should fix it.

The confirmation symbol is a green check mark, since you updated after initial check you need to press build & export all disks to get it, should take a few seconds.

Small issue but I believe should be an easy fix, if when doing multiple checks or builds you change WebGUI page to e.g. main, and some disks had finished, when you go back to plugin page finished disks appear as aborted, see screens for better example, latest version 2016.01.05a used.

Probably a good idea to add to initial page that to preclear disks on an empty server disk.cfg has to exist, quick and easy way is to do a new config, this will create a new disk.cfg even if there are no disks assign to the array.

It probably DOA, mine are Dell H310 but there’s a led regularly flashing from power on and the heatsink will feel warm to the touch after 30 seconds.

That was it, thanks. One little suggestion if you don’t mind, at the moment when the check is finish it only displays “Check finished”, I’d like to request more info, for example you could take the info displayed during a check and modify it a litle to display at the end: Check finished, processed xxx of xxx files. Skipped: 0 files. Found: 0 mismatches, 0 corruptions I believe you’re taking a simpler approach and just notifying in case of errors but I would feel better having more info of what was actually checked in the end.

OK, believe I found something, check works on hashs done with BLAKE2, fails on hashs done with MD5 or SHA256.

Select a disk and press build, wait for build to finish. Select same disk and press check. I even tried with simpler file names just in case and got same result: disk4.export.hash fe893c472a452790f85f7bb720fea856 */mnt/disk4/test/1.mp4 b2acaee8d6152c5fbaf27ad1abca9074 */mnt/disk4/test/2.mp4 61b6e399a721af7aa6ee53e55e5e05b8 */mnt/disk4/test/3.mp4 4e66607e129855e777f68c7c40e128cd */mnt/disk4/test/4.mp4 52b644ea8298b38b2891c7d89339db90 */mnt/disk4/test/5.mp4 syslog Jan 3 10:39:43 Testv6 bunker: added 5 files from /mnt/disk4 with mask *. Duration: 00:00:06 Average speed: 329. MB/s Jan 3 10:39:49 Testv6 bunker: warning: is missing Jan 3 10:39:49 Testv6 bunker: warning: is missing Jan 3 10:39:49 Testv6 bunker: warning: is missing Jan 3 10:39:49 Testv6 bunker: warning: is missing Jan 3 10:39:49 Testv6 bunker: warning: is missing

Updated but still not working, I’m only using new format only, from a quick test: Exported file: fe893c472a452790f85f7bb720fea856 */mnt/disk4/test/james.corden.2015.11.12.saoirse.ronan.hdtv.x264-crooks.mp4 b2acaee8d6152c5fbaf27ad1abca9074 */mnt/disk4/test/james.corden.2015.11.16.natalie.dormer.hdtv.x264-crooks.mp4 61b6e399a721af7aa6ee53e55e5e05b8 */mnt/disk4/test/james.corden.2015.11.17.ethan.hawke.hdtv.x264-crooks.mp4 4e66607e129855e777f68c7c40e128cd */mnt/disk4/test/james.corden.2015.11.18.anthony.mackie.hdtv.x264-crooks.mp4 52b644ea8298b38b2891c7d89339db90 */mnt/disk4/test/james.corden.2015.11.19.bryan.cranston.hdtv.x264-crooks.mp4 From syslog after pressing check: Jan 3 10:22:53 Testv6 bunker: warning: 5.11.12.saoirse.ronan.hdtv.x264-crooks.mp4 is missing Jan 3 10:22:53 Testv6 bunker: warning: 5.11.16.natalie.dormer.hdtv.x264-crooks.mp4 is missing Jan 3 10:22:53 Testv6 bunker: warning: 5.11.17.ethan.hawke.hdtv.x264-crooks.mp4 is missing Jan 3 10:22:53 Testv6 bunker: warning: 5.11.18.anthony.mackie.hdtv.x264-crooks.mp4 is missing Jan 3 10:22:53 Testv6 bunker: warning: 5.11.19.bryan.cranston.hdtv.x264-crooks.mp4 is missing

There appears to be something wrong in latest version, build works ok but trying to run a check completes without any file being verified, syslog fills with file not found errors, file names on log are incomplete, probably the cause of the problem.